On 12/28/2012 4:52 PM, Ralf Baechle wrote:
> On Thu, Dec 27, 2012 at 11:27:37PM -0500, Joshua Kinard wrote:
>> Has anyone run into an unhandled kernel unaligned access under 3.7.1? I've
>> triggered it twice w/ network I/O on an SGI IP32 machine, however, the stack
>> trace does not appear to be specific to any of IP32's own drivers. 3.6.7
>> was very stable, and the two oopses I've triggered so far both happened
>> under 3.7.1.
>> It looks like the culprit is in sk_stream_alloc_skb or tcp_sendmsg, however,
>> I have little experience in the higher-level networking stack within Linux
>> and wanted to see if anyone else has triggered this on other MIPS systems.
>> Seems to happen when I am logged in via SSH (on IPv6) and generating a burst
>> of console output.
>> Unhandled kernel unaligned access[#3]:
>> Cpu 0
>> $ 0 : 0000000000000000 0000000000000010 0000000000000000 bfffff005671271c
>> $ 4 : 0000000000000000 0000000000000000 0000000000000000 0000000000000000
>> $ 8 : 980000005c24e000 0000000000000000 980000005c24e000 00000000000000cc
>> $12 : ffffffff9001fce1 000000001000001e fffffffffffff000 000000000000001f
>> $16 : 980000005c00fa40 ffffffffde0300b8 ffffff0000000000 0000000000000005
>> $20 : 000000007f875700 00000000000005a8 0000000000000008 0000000000000005
>> $24 : 0000000000000001 00000000000003f0
>> $28 : 980000005c00c000 980000005c00fa10 0000000000000000 ffffffff800059a0
>> Hi : 0000000007a11c93
>> Lo : b645a1cac992645e
>> epc : ffffffff8000b700 do_ade+0x1b0/0x480
>> Tainted: G D
> This kernel has already oopsed before. Which means this oops message is
> pretty much worthless.
I'll try to recapture it over the weekend. This was the only output left in
the serial console buffer when I reattached, as I had the console
disconnected when it happened. Thanks!
"The past tempts us, the present confuses us, the future frightens us. And
our lives slip away, moment by moment, lost in that vast, terrible in-between."
--Emperor Turhan, Centauri Republic
Description: OpenPGP digital signature